usaga: Georgia State Highways

[ntallyn]

I have a handful of notes to still make my way through, as well as any comments on this thread beginning with #165.

I'm expecting to be able to push the last batch of updates over Thanksgiving weekend (including all of the items I've marked above as "local").

After I finish those updates, I will still need to go back and verify concurrencies/intersections among routes, and updating the concurrent US routes.

[CharlotteAllisonCDTG]

That sounds good.  Thank you for going through the updates and thread.

[yakra]

GA21:
GA21Spr -> GA21SprGar
GA21Spr_Spr -> GA21SprSpr

US40:	US40BusWhi
US73:	US40BusWhi    US40BusTho

Distinguish two different same-bannered same-numbered routes as needed with the 3-letter city abbreviations.

[ntallyn]

GA15 has broken concurrencies near Tallulah Gorge SP.
WilCon <-> +X036(GA15)
GA15Lp_S <-> US441BusCla_N

Are GA19 & US41BusMac concurrent between I-75 and the Georgia Ave / College St intersection? They're plotted differently.

Fixed the US highways locally for both of these. The one in Macon is quite a mess with the one-way couplet being several blocks wide and making a turn in the middle of it.

[ntallyn]

US 41 and SR 3: GA3ConMan -> GA3ConMar
US 41 and SR 3: requesting a point at http://www.openstreetmap.org/?lat=33.88871&lon=-84.47428 for WinRidPkwy
SR 120: requesting a point at http://www.openstreetmap.org/?lat=33.969417&lon=-84.489090 for ELakePkwy (I turned off onto Robinson Road, but that name is ambiguous since it returns to SR 120)

Local changes for now.

Fixed the spelling error.
Added WinRidPkwy (only because the similar AkeMillRd is already there on the other side of I-285).
Did not add ELakePkwy, as it doesn't really do much. It's only 1.5 miles between existing points already.

[ntallyn]

Just took a look at GA90. Many of the hidden shaping points are at or near road junctions, and should be converted to visible points.
"Prefer an intersection to act as a shaping point location wherever possible. Shaping points that coincide with intersections should be added as normal, visible waypoints labeled in the usual way."

Significantly updated (locally). Made matching changes to GA 11, 96, and 127 where concurrent.

[ntallyn]

US27:
GA100Ced_S -> GA100_Ced: this is a town suffix for disambiguation purposes, rather than a city abbreviation for one of multiple bannerless GA100 routes. With no other GA100_Ced point to disambiguate from (having US278_W instead), no need for the additional 'S' at the end.
GA100_End style labels have been deprecated a while; GA100/114, as GA1 has it, would work.

GA100:
US27Ced_S -> US27_Ced
US27 -> US27_Sum?

GA1: GA100Ced_S -> GA100_Ced
GA53: GA211_WinN -> GA211_Win
GA211: GA53Win_W -> GA53_Win

Similar labeling may be a wider issue.

One especially tricky one I noticed:
US1: GA46/86 http://www.openstreetmap.org/?lat=32.3692444&lon=-82.3236297
GA46: US1_S http://www.openstreetmap.org/?lat=32.3692444&lon=-82.3236297
These both use 7 decimal places while GA4 uses the standard 6, breaking the junction & concurrency.
GA4: GA46/86 http://www.openstreetmap.org/?lat=32.369244&lon=-82.323630

Fixed all of these above (as well as a couple of non-concurrency points on US1 and GA86 that had 7 decimal places.

There are some HIDDEN_JUNCTION data errors: http://travelmapping.net/devel/datacheck.php?rg=GA

A quick look at US27 & GA1 in side-by-side browser tabs shows lots of broken concurrencies. Some of these cases will be HIDDEN_JUNCTION, some VISIBLE_HIDDEN_COLOC, some NMPs, and some plain old points included in one route but not another. The HDX is a great help in checking concurrencies.

That will be part of the next batch of cleanup. I still have to learn how to use the various tools beyond the waypoint editor.

[ntallyn]

GA 28:
Should GA 28's southern segment remain in the system, since part of it in Augusta is a freeway (John C. Calhoun Expressway)?

GA 29/243:
With GA 243 being totally decommissioned, the portion that was concurrent with US 441 Bus (Milledgeville) has been redesignated as GA 29 Bus (Milledgeville).

GA 35:
GA 35, which is completely concurrent with US 319, has two special routes: GA 35 Conn (Thomasville) and GA 35 Loop (Tifton). Both are missing from the highway browser.

GA 73:
In the Statesboro area, GA 73 Byp (Statesboro) is missing from the highway browser. It is indeed a signed highway that is totally concurrent with US 301 Byp (Statesboro). Its southern portion is also concurrent with GA 67 Byp (Statesboro).

GA 540:
GA 540 is still missing from the highway browser.

GA28: If it's signed as GA28, it will remain in the system.
GA29Bus (Milledgeville): to be added.
GA243: Still signed along the Fall Line Freeway. New routing updated locally.
GA35 routes: to be added.
GA73Byp (Statesboro): to be added.
GA540: to be added.

[ntallyn]

GA21:
GA21Spr -> GA21SprGar
GA21Spr_Spr -> GA21SprSpr

Fixed locally.

[ntallyn]

I've made it through all of the comments above and made changes locally. Here's what's left:

GA11Byp (Cleveland): add
GA12Trk/US278Trk (Madison): add?
GA29Bus (Milledgeville): add?
GA32Trk (Douglas): add
GA35Con (Thomasville): add
GA35Lp (Tifton): add
GA37Trk (Camilla): add
GA45Trk (Colquitt): add
GA62Trk (Arlington): add
GA73: update routing
GA73Byp (Statesboro): add
GA75Con (Cleveland/White Co.): add
GA86 (Oak Park): update
GA91Trk (Colquitt): add
GA92: end location at Griffin? Probably leaving this one as-is for now.
GA154ConAtl: add?
GA158Trk (Douglas): add
GA166: update
GA216Trk (Arlington): add
GA246: split
GA540: add

US1 (Oak Park): update
US278Trk (Madison): add?

Gray Bypass (all routes): update

I hope to push all of the above before the end of the weekend. Then the hard work begins:

GA route concurrency/intersection checks
US route updates to reflect GA route updates

[ntallyn]

GA29Bus (Milledgeville): add?

Based on this turn sign https://www.google.com/maps/@33.0828522,-83.2338621,3a,75y,307.67h,89.07t/data=!3m6!1e1!3m4!1sUQcXlQH0uKlu9aGRDsSM5Q!2e0!7i13312!8i6656 (and one at the previous northbound turn), I'm going to add GA29BusMid for the entire route of US441BusMil, rather than ending it where GA243 used to end. Unfortunately, the signage goes cold after that, but this would also allow it to meet back up with GA29 on the north side of town.

[ntallyn]

GA32Trk (Douglas): add

I can find no signs along the presumptive route on GMSV. Nor are there any signs for Truck 32 heading east towards Douglas. The only signs I can find are approaching the loop around Douglas from the various other routes (including westbound 32). I also don't recall seeing any along the western side of town (following 206) when I was there last month.

Wanting to leave this out for now...

[ntallyn]

Pull request submitted:
https://github.com/TravelMapping/HighwayData/pull/3367

But there are conflicts (in the usausb.csv files). So obviously I'm doing something wrong. I figured when the system fetched the origin branch, it would update my master branch, as it does when I do it at work (though that is git, not github), and notify me of the conflict at that time. But that certainly isn't the case, as every file that I haven't changed has the time stamp of when I created the branch in the first place.

[yakra]

US27:
GA100Ced_S -> GA100_Ced: this is a town suffix for disambiguation purposes, rather than a city abbreviation for one of multiple bannerless GA100 routes.

Ack, lemme try to be more clear on this.
The abbreviation being tacked directly onto the end of a route designation (no underscore) implies there's a route by that .list name.
Whereas when there is a suffix, what comes before it would be the .list name (or just the route+banner, whatever's necessary to get rid of ambiguity).
Suffixes added to disambiguate otherwise-identical labels (when we're not using an abbrev) come after an underscore. Compare the _A _B _C _N _S _E _W suffixes that perform the same function.
It's the subtle difference between this rule and this one. In this case, the latter applies.

Thus:
US27 should have GA100_Ced, not GA100Ced
GA1 should have GA100_Ced, not GA100Ced
GA100 should have US27_Ced, not US27Ced. US27_Sum looks good.
GA211: GA53_Win looks good.

GA8ConVil:
http://205.209.84.174/devel/manual/wayptlabels.php#parentonly
This may show up on other bannered (spr/con/etc.) routes.

GA29BusMil:
AllMemDrSW -> AllMemDr
VinHwySE -> VinHwy (http://205.209.84.174/devel/manual/wayptlabels.php#dropdirection)
MLKingJrDr -> MarKingDr or MLKingJDr (http://205.209.84.174/devel/manual/wayptlabels.php#truncate)

GA35ConTho: GA35Con -> GA35

GA53:
There is no junction at GA1Lp_A.
GA211_Win label looks right

US19/129:
Missing an interchange at what's in the HB as US129TrkBla (which I assume no longer exists as such?); labeling might be tricky.

US23:
OldGA361 -> ArkRd or BassRd
OldBetRd is off (coords on GA42 are good)
GEHBMemHwy -> HoodAve or AnvBloRd (also affects GA42)
OldUS441Lak_S's coords are off. Realigned intersection? GA15 has better coords here. Should become OldUS441_S, unless the road at the actual intersection is determined to be part of Terrora Circle.
OldUS441Lak_N would become OldUS441_N or OldUS441 as the case may be.

US76:
US129Trk_S: Is this still a US129Trk route?

US129:
GA2_E -> US76_E
US76 +US76_W -> US76_W +US76

US411:
There is no junction at GA1Lp_A.

US441:
Many of the OldUS441 points have modern stable road names to use instead.
For example, OSM shows the first one as Bacon St.
The remaining ones will need their city suffixes shifted to after the underscore (An old route will not be in the database and will by definition not have an abbrev). Most of those with a directional suffix in addition to the city will need the directional character dropped, unless the city alone is insufficient to disambiguate.

US129TrkBla:  Does this still exist?

A lot of the rest of what I saw while skimming thru the changes looks good:
• Replacing OldFoo labels with modern stable road names like GriAve.
• Replacing hidden shaping points with visible points, particularly on GA90. I didn't look in exhaustive detail, but this probably goes a ways toward fixing broken concurrencies.

[yakra]

I've copied over the changes from #3367 into my fork, fixing the broken CSVs in the process.
Before we start discussion on merging it into the master branch, and how to best restore ntallyn:master to working order, a couple more notes...

US76:
Rather than move the in-use AltLabel US19/129_S to a new location and mess up stats & maps for those using it, the new point could be called US19_S, allowing us to leave the existing AltLabel in place. Probably best for US19/129_N (still in the HB as US19/129) to become US19_N in the process.

GA1LpRow needs a shaping point S of BlaBluRd

GA21AltPor:
missing Grange Rd interchange
SonDixInt -> JimDelPkwy (While Sonny Dixon Int does appear to be an actual road name, it's part of GA21Alt itself, not what GA21Alt intersects with.
GA25 -> GA21_N


I can get this stuff ready to merge back in, but would like to address some of these items from my last couple posts in the process.
@ntallyn, once you go to merge it back into your master branch, there will be merge conflicts. Rather than poke the git gods with a stick & try to resolve them, I'm thinking that resetting your master branch to the state it was in before your last commit may be the safer option. It's a clunky process, a bit roundabout. I can post instructions.